CT Terminology

How do we describe abnormalities on CT?

When describing findings on CT, we use words that refer to how dense the abnormality is in relation to a reference structure. Reference structures include bone, gray matter, white matter, CSF, fat, air etc.

Hypodense

Hypodense (less dense): If an abnormality is less dense than the reference structure, we would describe it as hypodense.

In the image to the right the solid arrow points to an area that is hypodense relative to the adjacent gray and white matter. This is an area of encephalomacia, or "softening" of the brain tissue, due to a previous infarction.

Isodense

Isodense (the same density): If an abnormality is the same density as the reference structure, we would describe it as isodense.

The solid arrow in the image above points to the supracellar cistern, a normal structure, which is filled with CSF. The dashed arrow points to an area that is isodense to CSF. This is a megacisterna magna, a normal structural variant.

Hyperdense

Hyperdense (more dense): If an abnormality is bright (white) on CT , we describe it as hyperdense.

On the image above, both arrows point to areas of hyperdensity. The solid arrow points to an area of right frontal and parietal intraparenchymal hemorrhage. The dashed arrow points to blood in the left lateral ventricle.


How is density measured on CT?

Radiodensity on CT is measured in Hounsfield Units (HU). HU range from -1000 to +1000. By definition water (CSF) = 0. Air is -1000 because it is the least dense structure. Bone is the most dense and measures +1000. Fat is less dense than water and therefore measures -100. Brain parenchyma is more dense than water and ranges from +20 to +40. White matter is less dense than gray matter due to the fat within the myelin within the white matter. Acute blood is bright on CT and measures + 55 to +75 HU. Calcification is more dense than blood and will measure in the low 100's. The image on the right is a graphic representation of different structures on CT with their relative HU values.